DESCRIPTION (adapted from the Abstract): The proposed project is a five- year study to analyze New Jersey (NJ) surveillance data on tuberculosis (TB) and AIDS, given that the results will generalize to other mixed urban-suburban areas of the United State, to further our understanding of the determinants of TB incidence and transmission during the AIDS era. The first objective of the research plan is to examine the impact of local conditions such as AIDS, foreign birth, and residential crowding on TB rates in New Jersey. A Poisson model incorporating spatial correlation will be used to construct estimates of local TB rates, overall, by drug resistance pattern and by HIV status of the subjects for 1996-1998. A spatio-temporal model of TB rates will also be constructed. The second objective is to contribute to a better understanding of the extent to which TV cases are due to recent transmission of the Mycobacterium tuberculosis by: estimating the number of people who are highly susceptible through construction of a spatial model of AIDS incidence; and predicting the number of TB causes due to recent transmission by using a mathematic model of the number of TB cases due to recent transmission by using a mathematic model of the number of TB causes the develop within two years of infection. Model inputs include the results of the spatial models of TB and AIDS, clinical and epidemiological features of TB, and demographic and vital statistics for New Jersey. The Latin Hypercue Sampling technique will be used to explore the sensitivity of the results to model inputs. The third objective will be used to explore the sensitivity of the results to model inputs. The third objective will be to validate the theory that clustered M. tuberculosis strains as determined by restriction fragment length polymorphism types are indicative of recently transmitted infection. This will be achieved by comparing mathematical model based prediction of the number of cases due to recent infection to predictions based on cluster analysis.